{"path":"research/non-zero-sum-economics-and-civilizational-cooperation.md","content":"# Non-Zero-Sum Economics, Civilizational Cooperation, and Deliberus\n\n**Date**: April 6, 2026\n**Purpose**: Deep academic exploration of how structured deliberation infrastructure — Deliberus specifically — enables humanity to discover and pursue non-zero-sum games that are currently invisible due to semantic confusion, information asymmetry, false polarization, and tribal epistemics. Extensive online research weaving together game theory, institutional economics, cooperative AI, deliberative democracy, anti-rivalrous information economics, David Deutsch's epistemology, and Robert Wright's Nonzero thesis.\n\n---\n\n## 1. The Core Thesis: Structured Reasoning Infrastructure as a Cooperation Multiplier\n\nDeliberus's civilizational vision — \"weaving our minds together\" — has an economic dimension that deserves rigorous treatment. The claim, made explicit:\n\n> **Much of what appears to be zero-sum conflict is actually non-zero-sum opportunity obscured by semantic confusion, information asymmetry, and broken sensemaking. A system that makes reasoning structure transparent, contested definitions visible, and shared premises discoverable would systematically reveal cooperative possibilities that are currently invisible.**\n\nThis is not naively optimistic. It does not claim all conflict is illusory. Žižek's parallax claims remain — some contradictions are irreducible, reflecting genuine structural antagonisms. But the proportion of disagreement that is genuinely irreducible versus semantically confused versus informationally asymmetric becomes *measurable* once reasoning is decomposed. And research consistently suggests that proportion skews heavily toward the resolvable end.\n\nThe thesis has three legs:\n\n1. **Non-zero-sum games are the engine of civilizational progress** (Wright, Deutsch, evolutionary theory)\n2. **Most of what blocks cooperative discovery is epistemic, not material** (false polarization, semantic confusion, narrative warfare)\n3. **Deliberus's specific architecture — decomposition, scheme classification, contested concept detection, bridging signals — directly addresses those epistemic blockers**\n\n---\n\n## 2. Robert Wright's Nonzero: The Directional Arrow\n\nRobert Wright's *Nonzero: The Logic of Human Destiny* (1999) provides the macro-historical frame. His central argument: both biological and cultural evolution are shaped by \"non-zero-sumness\" — the prospect of creating new interactions that produce mutual benefit. Each rung up the civilizational ladder increases both the amount and scope of non-zero-sum relationships.\n\n### The argument in brief\n\nNatural selection produces increasing complexity because organisms that discover cooperative arrangements — from mitochondrial symbiosis to multicellular coordination to social species to civilizations — outcompete those that don't. Information processing capacity is the key enabler: the better organisms (and their networks) get at processing information, the more non-zero-sum games they can identify and coordinate.\n\nWright traces this through human history: hunter-gatherer bands → chiefdoms → states → empires → global trade networks. At each level, the scope of non-zero-sum interaction expands. Trade is the paradigmatic example: voluntary exchange is positive-sum by definition (both parties prefer post-trade to pre-trade states), and its scope has expanded monotonically across human history.\n\n### Where Wright meets Deliberus\n\nWright's thesis implies that **the binding constraint on cooperation is information processing capacity** — specifically, the ability to identify, communicate, and coordinate around non-zero-sum opportunities. This is exactly Deliberus's operating domain.\n\nFredrik's 2012 formulation captures this precisely:\n\n> \"I'd like to find some algorithm for humanity to reason and make decisions together in a way that combines the democratic elements and general efficiency of a decentralized system like markets with the discerning force of logic.\"\n\nMarkets aggregate information via prices — dumb signals that tell you THAT resources should flow somewhere, not WHY. Hayek's insight was that this is efficient because individual knowledge is distributed and hard to centralize. But Wright's framework suggests the next evolutionary step is making the *argumentative layer* transparent — the reasoning behind resource allocation decisions, the evidence, the trade-offs, the value premises. This is what Deliberus does: make the WHY visible alongside the WHAT.\n\nWright's evolutionary frame also maps onto the Leverage Research notes (2012):\n\n> \"Evolutionary biology suggests next big thing is collaboration on a higher level of organization. We've seen this again and again, with the birth of multicellular life, culture, civilization, etc.\"\n\nEach major evolutionary transition enabled collaboration at a scale impossible at the previous level. Deliberus is proposed as infrastructure for the next transition: from civilizations with implicit reasoning to civilizations with explicit, navigable, structured reasoning — \"likened to the advent of language itself.\"\n\n### The information-theoretic connection\n\nWright explicitly links civilizational progress to information processing capacity. Deliberus is an information processing system for reasoning — it takes unstructured opinion and produces structured argument graphs with explicit premises, evidence links, scheme classifications, and contested concepts. This is literally an increase in the information processing capacity available for collective reasoning, which Wright's framework predicts should unlock new non-zero-sum opportunities.\n\n**Key source**: Wright, R. (1999). *Nonzero: The Logic of Human Destiny*. Pantheon. Also: [TED Talk: \"Progress is not a zero-sum game\"](https://www.ted.com/talks/robert_wright_progress_is_not_a_zero_sum_game).\n\n**2026 continuation — Wright follows his own arrow into the AI era.** Wright's \"The God Test\" (2026) extends this exact framework to an AI-augmented noosphere (a \"brain of brains of brains\") and restates the binding constraint for the AI era: governance is gated *psychologically*, on cognitive empathy — \"understanding adversaries' perspectives well enough to play non-zero-sum games with them.\" The 1999 constraint (information-processing capacity) and the 2026 constraint (cognitive empathy) are the same claim in different dress: the bottleneck is the cost of understanding the other party well enough to find the positive-sum move. His proposed instrument, though, is conversation with LLM companions — which his own market-selection argument (markets select for selectively-honest, approval-currying agents) undermines unless the mediation is measured; the connective tissue his global brain leaves unspecified is the slot this project builds for. Note also the inoculation the deeper read supplies: reviewers attack the book's directionality-of-evolution premise, but nothing borrowed here depends on it — the binding-constraint claim stands on transaction-cost grounds alone. Full read: [wright-god-test-noosphere-and-deliberus.md](wright-god-test-noosphere-and-deliberus.md).\n\n---\n\n## 3. David Deutsch and The Beginning of Infinity: Knowledge, Optimism, and Error-Correcting Institutions\n\nDavid Deutsch's *The Beginning of Infinity* (2011) provides the epistemological foundation that Wright's historical narrative lacks. Where Wright says \"non-zero-sum games drive progress,\" Deutsch explains *why* progress is possible at all and *what kind of institutions* sustain it.\n\n### The principle of optimism\n\nDeutsch defines optimism not as a feeling but as a consequence of epistemology: **\"All evils are caused by insufficient knowledge.\"** Problems are soluble — not because we already have the solutions, but because there is no known law of physics that prevents us from finding them. The critical requirement is not intelligence or resources but *institutions that enable error correction*.\n\nThis is a profound reframing for Deliberus. The question is not \"can humanity cooperate better?\" but \"what institutions would enable the discovery and correction of errors in our collective reasoning?\" Deliberus is proposed as exactly such an institution.\n\n### Good explanations and non-zero-sum discovery\n\nDeutsch's criterion for knowledge is the *good explanation* — hard to vary while still accounting for the phenomena. A good explanation constrains; a bad one can be adapted to fit anything.\n\nFor Deliberus, this provides a quality criterion for the argument graph: **a decomposition is better when it constrains the claim and clarifies why it stands or falls.** A distinction is better when it survives criticism and changes the reasoning, not when it is just another rhetorical restatement. The graph should gradually favor structures that are harder to vary while preserving explanatory force.\n\nApplied to non-zero-sum discovery: many apparent zero-sum conflicts persist because the explanations people use for their disagreements are *bad explanations* — easy to vary, vague enough to absorb any evidence. \"They disagree with us because they're ignorant/evil/tribal\" is a terrible explanation. \"They disagree because they're using a different operative definition of 'fairness' and weighting empirical evidence about incentive effects differently\" is a much better explanation — and one that immediately suggests cooperative possibilities.\n\nDeliberus's extraction pipeline, by forcing decomposition into atomic claims with explicit type classification (empirical, definitional, normative, inference), structurally pushes toward better explanations of disagreement. The contested concept detection (\"you mean freedom-from-coercion; they mean freedom-from-deprivation\") makes bad explanations of disagreement *visibly* bad.\n\n### Conjecture and criticism, not justification\n\nDeutsch's Popperian epistemology — knowledge grows through conjecture and criticism, not through justification from foundations — maps directly onto the \"No Copout Axioms\" principle. The platform is not a machine for certifying final foundations. It is infrastructure for exposing assumptions, surfacing criticisms, and letting better structures emerge. A claim is not made valuable by being insulated from attack but by surviving better attacks.\n\nThis has direct economic implications. In a Deutschian framework, **the reason humanity doesn't cooperate more is not that cooperation is impossible but that we haven't yet created the knowledge needed to see the cooperative possibilities.** The non-zero-sum games are there; we just can't see them through the fog of semantic confusion and narrative warfare. Error-correcting institutions — institutions that make criticism visible and productive — are the mechanism by which we discover them.\n\n### Wealth as transformative capability\n\nDeutsch redefines wealth: \"It is the set of all transformations that you are capable of bringing about.\" Knowledge is the multiplier — a given individual's wealth is proportional to the knowledge available to them far more than to the physical resources they control.\n\nThis maps onto the Deliberus civilizational graph as a *wealth creation engine*: by making the structure of human reasoning transparent, navigable, and improvable, the graph increases the transformative capability of every person who can access it. A citizen who can trace a policy recommendation from conclusion through every logical step to its value premises and empirical evidence has more transformative capability than one who can only vote on a slogan.\n\n### Error-correcting institutions and open societies\n\nDeutsch, following Popper, argues that the decisive difference between societies that progress and those that stagnate is the presence of **error-correcting institutions** — social and political structures that make it possible to detect mistakes, remove bad policies, and recover from disasters without violence.\n\n> \"The only rational policy is to judge institutions, plans and ways of life according to how good they are at correcting mistakes: removing bad policies and leaders, superseding bad explanations, and recovering from disasters.\"\n\nDemocracy is not valuable because it produces the best leaders (it often doesn't) but because it provides a mechanism for removing bad ones without bloodshed. Science is not valuable because scientists are always right but because the institution of peer review and replication enables error correction.\n\nDeliberus, in this framework, is a **proposed error-correcting institution for collective reasoning.** The argument graph makes errors in reasoning visible. The sorry markers make incompleteness visible. The critical questions generated from Walton's schemes force examination of assumptions. The bridging signal surfaces reasoning that survives cross-group scrutiny. Each of these is an error-correction mechanism — not proving conclusions right, but making the structure of reasoning transparent enough that errors can be found and fixed.\n\n### Constructor theory and what's possible\n\nDeutsch's more recent work on constructor theory — a framework for physics in terms of what transformations are possible and what are impossible — has an intriguing meta-connection to Deliberus. Constructor theory asks: what *can* be caused to happen? Applied socially: what kinds of cooperation *can* be constructed given the right institutional design?\n\nThe question is not \"will humanity cooperate?\" but \"is there a constructor — an institutional arrangement — that reliably produces cooperation as output given disagreement as input?\" Deliberus is a proposed constructor for structured cooperation from unstructured disagreement.\n\n**Key sources**: Deutsch, D. (2011). *The Beginning of Infinity: Explanations That Transform the World*. Viking. [David Deutsch and Naval Ravikant — Tim Ferriss Show #662](https://tim.blog/2023/03/24/david-deutsch-naval-ravikant-transcript/). [The Popperian Podcast #1 — David Deutsch on Popper](https://www.jedleahenry.org/the-popperian-podcast/2020/11/29/the-popperian-podcast-1-david-deutsch-karl-popper-and-the-beginning-of-infinity).\n\n---\n\n## 4. Why We Miss Non-Zero-Sum Games: The Epistemic Blockers\n\nIf non-zero-sum opportunities are so abundant, why does so much of human interaction remain zero-sum or worse? Research identifies several mechanisms, each of which Deliberus's architecture directly addresses.\n\n### 4.1 False polarization: We overestimate our disagreements\n\nThe \"Hidden Tribes\" study (More in Common, 2018) found that **67% of Americans belong to an \"Exhausted Majority\"** — fatigued by polarization and eager for cooperation. Democrats and Republicans imagine almost twice as many political opponents hold \"extreme\" views as actually do. The \"Perception Gap\" research found that the more news people consume, the *larger* their misperception of the other side — media consumption actively worsens the accuracy of cross-partisan mental models.\n\nFalse polarization research (Lees & Cikara, 2021; Blatz & Mercier, 2018) identifies three cognitive mechanisms: categorical thinking (binary us/them), oversimplification (reducing complex positions to caricatures), and emotional amplification (affective polarization driving perception of issue polarization). People overestimate the extremity of opponents' ideologies while underestimating their certainty.\n\n**Deliberus connection**: The argument graph makes actual positions visible at granular resolution. Instead of \"they believe X\" (where X is an imagined extreme), users see the specific premises, evidence, and reasoning structure. The contested concept detection shows WHERE definitions diverge — often revealing that the \"extreme\" position is actually a different definition of the same term, not a different value. The bridging signal explicitly surfaces claims where the REASONING quality is high despite conclusion disagreement — the \"I can't dismiss this\" moment that false polarization makes invisible.\n\n### 4.2 Semantic confusion: Same words, different meanings\n\nThe Deliberus project has documented this extensively — the Swedish \"barnfattigdom\" (child poverty) debate where everyone agreed on facts but different operative definitions generated a public fight about nothing. Fredrik's 2012 Leverage Research notes identified this with remarkable precision:\n\n> \"Isomorphisms between paths to conclusion? Or just isomorphism between conclusions? Illusory agreement common/pervasive?\"\n\nResearch confirms this is pervasive. Disputes over \"freedom,\" \"fairness,\" \"equality,\" \"security,\" and \"rights\" routinely involve parties using the same word to mean different things. These are not genuine value disagreements — they are coordination failures caused by linguistic ambiguity.\n\n**Deliberus connection**: Contested concept detection is a first-class feature of the extraction pipeline. The system identifies terms with multiple operative definitions across worldviews and makes them explicit as definitional claim nodes. The moment a user sees \"oh, we were using the same word differently — THAT'S why we disagree\" is one of the most powerful the platform offers. This directly converts what appeared to be a zero-sum ideological conflict into a tractable definitional clarification task.\n\n### 4.3 Information asymmetry and narrative warfare\n\nSchmachtenberger's analysis of narrative warfare — deploying true-but-selectively-framed facts to support predetermined conclusions — describes a systematic mechanism for obscuring non-zero-sum possibilities. When each side sees only the evidence that supports their position, cooperative opportunities that require synthesizing evidence from both sides remain invisible.\n\nMarkets partially solve information asymmetry through price signals, but only for quantifiable goods. For reasoning about values, policies, and collective decisions, no equivalent aggregation mechanism exists at scale.\n\n**Deliberus connection**: The argument graph makes framing visible by decomposing claims into their constituent premises and evidence. When two narratives about the same issue are both extracted, the graph shows where they share premises, where they diverge, and what evidence each omits. The \"mapped perspectives, not truth\" framing explicitly resists narrative capture. The QBAF gradual semantics score each claim based on the quality of its evidence and the status of its critical questions — not on who said it or how many people believe it.\n\n### 4.4 Institutional failure to aggregate distributed knowledge\n\nHayek's insight was that knowledge is distributed and no central planner can aggregate it. Markets aggregate quantitative information (prices), but reasoning about complex social questions involves qualitative structure that prices cannot capture: \"Policy X works because of mechanism Y, but only under conditions Z, and it has side effect W that affects population P disproportionately.\"\n\nCurrent institutions for aggregating reasoning — legislatures, media, academia — each have structural limitations: legislatures compress to binary votes, media compresses to narratives, academia publishes in silos. None produces a persistent, structured, publicly navigable representation of the full reasoning landscape.\n\n**Deliberus connection**: The civilizational deliberation graph is proposed as a new institution for aggregating distributed reasoning — what the scientific method did for empirical inquiry, generalized to all domains of deliberation. Tom Boniecki's 2011 analogy is apt: \"a spreadsheet for knowledge rather than an expert system\" — the system doesn't tell you what to conclude; it gives you a structure for organizing reasoning that makes it inspectable and improvable.\n\n---\n\n## 5. Elinor Ostrom and the Governance of Epistemic Commons\n\nElinor Ostrom's Nobel Prize-winning work on governing the commons provides a rigorous framework for understanding Deliberus as an institutional design challenge.\n\n### Ostrom's design principles\n\nOstrom identified eight design principles for sustainable commons governance, derived from empirical fieldwork across hundreds of communities worldwide. Several map directly onto Deliberus's design:\n\n| Ostrom Principle | Deliberus Application |\n|---|---|\n| **Clearly defined boundaries** | Who contributes to the graph? How are quality standards maintained? The reputation system (calibration, argument quality, intellectual honesty) defines contribution boundaries |\n| **Congruence between rules and local conditions** | Progressive disclosure — different rules for different participation depth. Casual users vote; experts decompose; curators maintain canonical structure |\n| **Collective-choice arrangements** | \"The rules of the system are themselves deliberatable\" (Lean kernel analogy) — the community can argue about the platform's own foundations |\n| **Monitoring** | QBAF gradual semantics, calibration scores, cross-adversarial quality ratings — visible, algorithmic monitoring of contribution quality |\n| **Graduated sanctions** | Signal-weight gates (new accounts start at lower visibility), privilege unlocks tied to demonstrated epistemic behavior |\n| **Conflict-resolution mechanisms** | The argument graph IS the conflict resolution mechanism — structured decomposition of disagreements into their constituent premises |\n| **Nested enterprises** | Polycentric governance — topic-specific reputation, domain-specific calibration, local community norms within global graph structure |\n\n### The epistemic commons as anti-rival good\n\nSchmachtenberger's \"epistemic commons\" concept maps onto Ostrom's framework with an important extension: the deliberation graph is not merely non-rivalrous (my use doesn't diminish yours) but *anti-rivalrous* — it becomes MORE valuable as more people contribute. Every well-formulated argument makes the graph more useful for everyone, including those who disagree with it. Every identification of a contested concept or evidence gap improves the map for all participants.\n\nSteven Weber (2004) formalized this property: an anti-rival good meets the test of a public good (non-excludable, non-rivalrous) but additionally gains value from each new contributor. English as a language is an anti-rival good — the more speakers, the more useful for each speaker. The argument graph has the same property: the more worldviews represented, the more complete the map, the more useful for any individual navigator.\n\nThis is structurally different from social media (where engagement is rivalrous — your attention is taken from mine) and from debate (where winning is zero-sum). The graph IS the non-zero-sum game — contributing to it is positive-sum by construction.\n\n### Polycentric governance of knowledge\n\nOstrom's key insight was that neither pure market nor pure state governance works for commons — what works is *polycentric governance* with overlapping authorities at multiple scales. For Deliberus, this suggests:\n\n- **Local governance**: Topic communities manage their own quality standards and canonical claims\n- **Global governance**: Cross-topic structural rules (scheme classification, claim types) maintained by a small, auditable kernel (the Lean analogy)\n- **Emergent governance**: Worldview clusters emerge bottom-up from voting patterns, not imposed top-down from taxonomies\n\nThis is exactly the architecture described in the Lean analogies research: \"The social system IS the verification system. The machine checks structure; the community judges meaning.\"\n\n**Key source**: Ostrom, E. (1990). *Governing the Commons: The Evolution of Institutions for Collective Action*. Cambridge University Press. [Summary — Beyond Intractability](https://www.beyondintractability.org/bksum/ostrom-governing).\n\n---\n\n## 6. Cooperative AI and Mechanism Design for Non-Zero-Sum Discovery\n\n### Dafoe's cooperative AI framework\n\nAllan Dafoe's \"Open Problems in Cooperative AI\" (2020) — which catalyzed a $15 million foundation — identifies a set of fundamental challenges for AI systems that need to cooperate. The core insight: achieving cooperative behavior among intelligent agents in complex environments remains poorly understood, even when cooperation would benefit all parties.\n\nThe framework identifies five research areas: understanding (modeling others), communication (sharing information), commitment (binding agreements), institutions (rules and norms), and sociality (group identity and norms). Deliberus intersects primarily with *understanding* (making reasoning visible), *communication* (structured argument sharing), and *institutions* (the graph as cooperative infrastructure).\n\nA key finding from the broader cooperative AI literature: complex systems perspectives on reinforcement learning show that cooperation challenges like climate change mitigation and social dilemmas — where individual incentives don't align with collective interest — require explicit mechanisms for making collective welfare visible. The argument graph does this for reasoning: individual contributions improve collective understanding, and the system makes this improvement visible.\n\n### Mechanism design: incentivizing cooperation discovery\n\nMechanism design theory (Hurwicz, Maskin, Myerson — Nobel 2007) studies how to design rules for interactions so that self-interested participants produce socially desirable outcomes. The prediction markets research doc already covers LMSR and futarchy as specific mechanisms.\n\nFor non-zero-sum discovery specifically, the relevant mechanism design insight is: **cooperative outcomes often fail not because agents are selfish but because the information structure doesn't reveal the cooperative possibility.** The prisoner's dilemma is zero-sum only under conditions of non-communication and non-repetition. Add communication (agents can reason about each other's reasoning) and repetition (agents build track records), and cooperation becomes the dominant strategy.\n\nDeliberus adds a third dimension: *structured reasoning transparency*. Not just \"I can communicate with you\" but \"I can see the full structure of your reasoning — which premises you accept, which evidence you weight, which definitions you use — and you can see mine.\" This is a fundamentally richer information structure than either prices (markets) or votes (democracy) provide.\n\n### The Procaccia/Konya peacebuilding existence proof\n\nThe most striking empirical validation of structured deliberation producing cooperation in seemingly zero-sum settings: Procaccia, Konya et al. (FAccT 2025) used collective dialogues combined with AI-mediated bridging to find common ground between **Israeli and Palestinian peacebuilders during active conflict.**\n\n138 civil society peacebuilders participated — Israeli Jews, Palestinian citizens of Israel, and Palestinians from the West Bank and Gaza. The process used bridging-based ranking (structurally similar to Polis but enhanced with LLM synthesis) to identify \"bridging statements\" endorsed across group boundaries. The result: **a set of collective statements with at least 84% agreement from participants on each side.**\n\nThis is an *existence proof* that structured deliberation can surface non-zero-sum possibilities in the hardest possible case: parties with genuine, deep, material grievances, during active violent conflict. If bridging works there, the potential for non-zero-sum discovery in less extreme settings is vast.\n\nThe Deliberus architecture goes further than the Procaccia/Konya methodology in two ways: (1) it produces persistent, evolving argument graphs rather than one-off consensus statements, and (2) it decomposes reasoning structure (scheme classification, critical questions, contested concepts) rather than just clustering opinion statements. The bridging signal in Deliberus — `disagreement_factor × QBAF_strength` — captures not just \"both sides agree on this statement\" but \"both sides recognize this REASONING as structurally sound despite disagreeing on the conclusion.\" This is a richer signal.\n\n**Key source**: Procaccia, A., Konya, L. et al. (2025). \"Using Collective Dialogues and AI to Find Common Ground Between Israeli and Palestinian Peacebuilders.\" FAccT 2025. [arXiv:2503.01769](https://arxiv.org/abs/2503.01769).\n\n---\n\n## 7. Deliberative Democracy: Empirical Evidence for Cooperation Through Structured Reasoning\n\n### Fishkin's deliberative polling\n\nJames Fishkin's decades of deliberative polling research provides the strongest empirical foundation for the claim that structured reasoning produces cooperative outcomes. Key finding: **~70% of participants change positions on at least one major issue** when exposed to balanced, structured arguments from both sides. The key is exposure to the *best versions* of opposing arguments, not strawmen.\n\nFishkin's \"Twelve Key Findings in Deliberative Democracy Research\" (American Academy of Arts and Sciences) establishes:\n\n1. Deliberation produces outcomes superior to other forms of democracy\n2. Participants show less partisanship and more sympathy with opposing views afterward\n3. Respect for evidence-based reasoning increases\n4. Commitment to decisions increases (because participants understand why)\n5. Broadly shared consensus emerges more frequently\n\nIn La Plata, Argentina, a representative deliberative process on transit policy showed trust in government increasing dramatically: disagreement with \"public officials care about what people like me think\" dropped from 60% to 20% after participation.\n\n### Citizens' assemblies and economic cooperation\n\nCitizens' assemblies — randomly selected groups of citizens deliberating on complex policy questions with expert input — have produced remarkably cooperative outcomes on issues typically considered zero-sum:\n\n- **Ireland's Citizens' Assembly** (2016-2018): Produced recommendations on abortion, climate, and aging that broke decades of political deadlock. The assembly recommended liberalizing abortion law, which passed in a referendum by 66%.\n- **France's Citizens' Convention on Climate** (2019-2020): 150 randomly selected citizens produced 149 proposals for climate action, many of which were adopted by the government.\n- **Taiwan's digital deliberation** (vTaiwan + Polis): Resolved contentious ride-hailing regulation through bridging-based consensus, creating policy that satisfied both taxi drivers and ride-hailing advocates.\n\nThe pattern across all cases: **structured deliberation with adequate information consistently discovers cooperative solutions that adversarial politics could not.** The non-zero-sum games were there all along; adversarial framing made them invisible.\n\n### What deliberation does that voting doesn't\n\nResearch consistently shows that deliberation produces *preference transformation*, not just preference aggregation. Participants don't just learn what others think — they *change what they themselves think* in response to reasoning they hadn't previously encountered. This is the mechanism by which hidden non-zero-sum games become visible: people discover that their opponents' reasoning addresses concerns they hadn't considered, and their own position was partly based on missing information.\n\nDeliberus automates and scales several aspects of this process: the extraction pipeline surfaces argument structure that would take hours of facilitated discussion to make explicit; the contested concept detection identifies definitional confusion that might never surface in unstructured deliberation; the bridging signal computationally identifies the most productive starting points for cross-group engagement.\n\n**Key source**: Fishkin, J. (2018). *Democracy When the People Are Thinking*. Oxford University Press. [Twelve Key Findings — American Academy of Arts and Sciences](https://www.amacad.org/publication/daedalus/twelve-key-findings-deliberative-democracy-research).\n\n---\n\n## 8. The Economic Theory of Non-Zero-Sum Reasoning\n\n### Markets aggregate prices; Deliberus aggregates reasoning\n\nThe standard economic argument for markets is Hayekian: distributed knowledge is aggregated through the price mechanism more efficiently than any central planner could achieve. This is correct for quantifiable goods with clearly defined property rights.\n\nBut many of the most important collective decisions — climate policy, AI governance, public health, urban planning — involve qualitative reasoning that prices cannot capture. \"Should we invest in nuclear or solar?\" is not answerable by a price signal alone; it requires weighing safety evidence, climate projections, economic models, political feasibility, equity considerations, and value premises about intergenerational responsibility.\n\nFor these decisions, the equivalent of a market's price aggregation is **argument structure aggregation** — making the full reasoning landscape visible so that decision-makers can identify where evidence is strong, where it's contested, where genuine value differences lie, and where apparent conflict dissolves under semantic analysis.\n\nDeliberus is proposed as this aggregation mechanism. The argument graph is to reasoning what the price system is to resource allocation: a mechanism for making distributed knowledge visible and actionable.\n\n### Positive-sum knowledge creation\n\nDeutsch's redefinition of wealth — \"the set of all transformations you are capable of bringing about\" — implies that **knowledge creation is inherently positive-sum.** Unlike physical resources, knowledge is non-rivalrous (my use doesn't diminish yours) and often anti-rivalrous (the more people use an explanation, the more refined and powerful it becomes).\n\nThe argument graph amplifies this: each extraction enriches the graph. The `@[simp]` flywheel (from Lean) means every well-vetted argument makes the system better at automatically connecting future arguments to existing ones. This is compound growth in collective reasoning capacity — the same positive-sum dynamic that makes markets and science engines of wealth creation, applied to the domain of values and policy.\n\n### Transaction costs and the Coase theorem\n\nRonald Coase's insight (1960) was that in a world without transaction costs, parties would always negotiate to the most efficient outcome regardless of initial property rights allocation. The reason they don't is transaction costs — the cost of identifying, negotiating, and enforcing agreements.\n\nApplied to deliberation: much zero-sum political conflict persists because the *transaction costs of mutual understanding* are too high. Understanding why your opponent holds their position — their evidence, definitions, values, reasoning structure — requires hours of patient dialogue. Most people don't invest this. The result: negotiations over policy start from mutual misunderstanding and settle at suboptimal compromises.\n\nDeliberus dramatically reduces the transaction costs of mutual understanding by making reasoning structure persistent, navigable, and granular. Instead of spending hours in dialogue to discover \"oh, they mean freedom-from-deprivation while I mean freedom-from-coercion,\" the user sees this in seconds via the contested concept page. The Coasian prediction: lowering these transaction costs should shift many apparent zero-sum conflicts toward cooperative outcomes.\n\n### Positive externalities of structured reasoning\n\nStructured reasoning has positive externalities that unstructured debate does not:\n\n1. **Deduplication**: \"We already mapped this argument; here's what we found; here's where it's still open.\" Prevents re-litigating settled premises.\n2. **Composability**: Arguments in the graph can be composed — if A supports B and B supports C, the system can compute the transitive strength. This is impossible with unstructured text.\n3. **Error detection**: Formal scheme classification + critical question generation + gap detection systematically surface errors that unstructured debate conceals.\n4. **Knowledge persistence**: Unlike conversations (ephemeral) or articles (static), the graph evolves — new evidence updates the structure, new arguments connect to existing ones.\n\nEach of these reduces the cost of collective reasoning, which Wright's framework predicts should unlock new cooperative possibilities.\n\n---\n\n## 9. The Convergence Thesis as Economic Prediction\n\nThe Deliberus convergence thesis — that values converge when decomposed far enough — has a striking economic implication: **if the convergence thesis holds, then the space of non-zero-sum games is far larger than it appears.**\n\n### Why convergence predicts cooperation\n\nIf most disagreement lives in the \"middle layers\" (definitions, cultural framing, empirical beliefs) rather than at the value bedrock, then most apparent zero-sum conflicts are actually misidentified. Two parties who appear to want incompatible things may actually want the same thing and disagree only about definitions or empirical mechanisms.\n\nThe MGE paper (Klingefjord, Lowe & Edelman, 2024) provides the strongest existing evidence: when values are structurally elicited through LLM-guided decomposition, **participants overwhelmingly converge on the directionality of value relationships.** Convergence is not imposed — it emerges from structured decomposition.\n\nThe economic prediction: a system that decomposes disagreements to sufficient depth will systematically discover cooperative possibilities invisible to surface-level debate. The non-zero-sum game was always there; the linguistic and cognitive compression of ordinary discourse kept it hidden.\n\n### The false polarization connection\n\nIf the Hidden Tribes research is right that 67% of Americans are an \"Exhausted Majority\" who overestimate partisan differences, and if false polarization research is right that people imagine twice as many opponents hold extreme views as actually do, then **the current level of political zero-sum competition is substantially above what the actual distribution of preferences would produce under full information.**\n\nStructured deliberation is a mechanism for correcting this misperception — not by telling people they agree (that would be patronizing and often wrong) but by showing them the granular structure of where they agree and where they don't.\n\n### Testing the convergence thesis economically\n\nThe convergence thesis generates a testable economic prediction: **communities that use structured deliberation tools should discover more cooperative solutions to resource allocation problems than communities that use adversarial debate.** This could be tested via randomized controlled trials comparing Deliberus-mediated policy discussions with standard public comment processes or town halls. The Procaccia/Konya study provides the methodological template.\n\n---\n\n## 10. Deutsch, Wright, and the Architecture of Open-Ended Progress\n\n### Error correction as the meta-game\n\nSynthesizing Deutsch and Wright: **the deepest non-zero-sum game is the meta-game of improving our collective ability to discover non-zero-sum games.** Error-correcting institutions compound — each improvement in our ability to detect and correct errors in collective reasoning makes us better at finding cooperative opportunities, which provides resources for further institutional improvement.\n\nDeutsch calls this \"the beginning of infinity\" — an open-ended process of knowledge creation with no inherent limits. Wright calls it the \"logic of human destiny\" — the directional arrow of increasing non-zero-sum scope.\n\nDeliberus is proposed as a concrete next step in this meta-game: an institution explicitly designed to make reasoning errors visible and correctable at civilizational scale.\n\n### Why pessimism about cooperation is usually wrong (Deutsch)\n\nDeutsch's argument against pessimism has direct relevance: **pessimistic predictions about the impossibility of cooperation typically assume a fixed level of knowledge.** \"Humanity will never agree on climate policy\" assumes the current level of mutual understanding is the maximum achievable. But if institutions can be improved — if better tools for making reasoning transparent can be built — then the level of mutual understanding can increase, and the space of possible cooperation expands.\n\nThis is not Panglossian optimism. It's a consequence of epistemology: problems are soluble, including the problem of insufficient cooperation. The question is whether we build the institutions that enable the solution.\n\n### The anti-zero-sum institution\n\nBoth Deutsch and Wright imply a design criterion for institutions: **good institutions convert zero-sum interactions into non-zero-sum ones.** Markets do this for resource allocation (converting zero-sum physical competition into positive-sum trade). Democracy does this for governance (converting zero-sum power struggles into peaceful transitions). Science does this for empirical knowledge (converting zero-sum prestige contests into positive-sum knowledge accumulation, at least in principle).\n\nDeliberus is proposed to do this for reasoning about values and policy — converting zero-sum ideological warfare into positive-sum deliberation where each contribution (even a wrong one) improves the collective map.\n\n---\n\n## 11. Open Questions and Research Frontier\n\n### Can the non-zero-sum discovery thesis be empirically tested?\n\nThe strongest version of the thesis — that structured deliberation systematically reveals cooperative possibilities invisible to adversarial debate — generates testable predictions:\n\n1. Communities using Deliberus-style decomposition should reach cooperative agreements more often than control groups using unstructured discussion\n2. The proportion of disagreements identified as \"semantic\" (same-word-different-meaning) by the contested concept detector should correlate with the proportion that are resolvable\n3. The bridging signal (`disagreement × QBAF_strength`) should predict which topics are most amenable to cooperative resolution\n\n### What is the economic value of reduced semantic confusion?\n\nIf the Swedish \"barnfattigdom\" debate was entirely semantic and wasted significant public discourse energy, what is the economic value of a system that identifies semantic confusion early? Could the cost savings from avoiding futile zero-sum debates over definitions alone justify the platform?\n\n### Does Deutsch's optimism survive Žižek's challenge?\n\nŽižek's parallax claims argue that some contradictions are irreducible — not resolvable by more information or better reasoning. If this is true for a significant fraction of political disagreement, then the non-zero-sum discovery thesis has hard limits. The convergence thesis doesn't require ALL disagreement to be resolvable — just enough to make the investment in better reasoning infrastructure worthwhile. The empirical question is: what fraction?\n\n### Can anti-rivalrous dynamics survive adversarial use?\n\nThe argument graph is anti-rivalrous when participants contribute in good faith. But Schmachtenberger's warning applies: sophisticated actors might learn to game the graph, contributing structurally sound but misleadingly framed arguments. Can the reputation system (calibration scoring, cross-adversarial quality ratings, intellectual honesty records) maintain anti-rivalrous dynamics under adversarial pressure?\n\n### How does the Lean analogy extend to economic institutions?\n\nIf Lean's `sorry`-driven collaboration enables 25 strangers to formalize a proof without trusting each other (Tao's PFR project), can analogous mechanisms enable strangers to cooperate on economic problems without trusting each other? The `sorry` model converts \"I don't have the answer\" from a weakness into an invitation. Could economic institutions be designed with explicit `sorry` markers — \"we don't know the optimal policy here; contribute your knowledge to fill in this gap\"?\n\n---\n\n## 12. Synthesis: How Deliberus Enables Non-Zero-Sum Discovery\n\n| Mechanism | What It Does | Non-Zero-Sum Impact |\n|---|---|---|\n| **Atomic claim decomposition** | Breaks vague disagreements into specific, addressable premises | Reveals that apparent conflicts are often about different sub-questions, many of which are non-zero-sum |\n| **Contested concept detection** | Identifies terms being used with different meanings | Converts semantic conflicts (zero-sum word fights) into definitional clarification tasks (cooperative) |\n| **Scheme classification + CQ generation** | Maps argument structure and generates critical questions | Makes reasoning quality assessable independently of agreement — prerequisite for bridging |\n| **Bridging signal** | Surfaces reasoning that both sides find structurally sound | Identifies the non-zero-sum territory where engagement is most productive |\n| **QBAF gradual semantics** | Scores claims by evidence quality, not popularity | Prevents cooperative reasoning from being drowned out by tribal signaling |\n| **Worldview filter** | Lets users see the graph from other perspectives | Reduces false polarization by making actual positions (not imagined extremes) visible |\n| **`@[simp]` flywheel** | Each contribution improves the system for everyone | Anti-rivalrous knowledge accumulation — the cooperation multiplier |\n| **Sorry markers** | Make incompleteness visible and inviting | Convert \"I don't know\" from zero-sum vulnerability into positive-sum contribution opportunity |\n| **Persistent graph** | Reasoning accumulates across time and contributors | Reduces transaction costs of mutual understanding — the Coasian precondition for cooperative outcomes |\n\n---\n\n## Conclusion: The Civilizational Argument\n\nThe argument, in its strongest form:\n\n1. **Non-zero-sum games drive civilizational progress** (Wright: the directional arrow of increasing cooperative scope)\n2. **Progress is unlimited in principle** (Deutsch: problems are soluble, knowledge growth is open-ended)\n3. **The binding constraint is knowledge — specifically, the ability to identify cooperative possibilities** (Deutsch: all evils are caused by insufficient knowledge)\n4. **Most apparent zero-sum conflicts contain hidden non-zero-sum structure** (false polarization research, convergence thesis, semantic confusion evidence)\n5. **Current institutions fail to surface this structure** (markets aggregate prices but not reasoning; media compresses to narratives; legislatures compress to binary votes)\n6. **Deliberus is designed to surface this structure** (decomposition, scheme classification, contested concepts, bridging signal, worldview filter)\n7. **Therefore, Deliberus — or something like it — is civilizational infrastructure for accelerating the discovery of cooperation**\n\nThis is the economic case for Deliberus, expressed in the language of game theory, institutional economics, and epistemology. It doesn't claim Deliberus will solve all conflict. It claims that a significant fraction of what looks like irreducible conflict is actually solvable-but-hidden, and that the architecture described here is designed to reveal it.\n\n> \"In order to figure out (and transition into) something better than capitalism, we need to build a collaborative sensemaking platform, Deliberus or something like it.\"\n> — Fredrik (conversation with Edvin, 2011)\n\nFifteen years later, the academic literature converges: the tools now exist, the theory supports it, and the empirical evidence — from Fishkin's deliberative polls to the Procaccia/Konya peacebuilding existence proof — validates the core premise. The non-zero-sum games are there. We just need the infrastructure to see them.\n\n---\n\n## 13. Deep Research Findings: Satellite Documents Integration\n\nFive parallel research agents produced ~170KB of additional findings. This section integrates the most significant data points and insights. Full details in the companion documents listed in Sources.\n\n### 13.1 Quantified False Polarization (from trust-economics research)\n\nThe most striking numbers for the Deliberus thesis:\n\n- **Republicans think 50% of Democrats believe \"most police are bad people\"** — actual: 15%. Democrats think only 30% of Republicans support reasonable gun control — actual: ~70%. The perception gap is roughly 2x across the board.\n- **More education worsens accuracy**: Postgraduate Democrats are 3x *less accurate* about Republican positions than those without high school diplomas. This is identity-protective cognition at work — sophistication enables better motivated reasoning, not better calibration.\n- **Kahan's motivated numeracy**: High-numeracy partisans are 45% more likely to misread data when the correct answer threatens their identity (vs 25% for low-numeracy). Intelligence amplifies polarization. But *science curiosity* (distinct from science knowledge) counteracts this — a crucial design insight: Deliberus should cultivate curiosity, not just present information.\n- **Global cost of unnecessary conflict**: $19.97 trillion/year (11.6% of global GDP). U.S. trust in government: 73% (1958) → 22% (2024). A single 43-day government shutdown cost $11 billion.\n- **Moral universalism confirmed**: Curry et al. (2019) found seven moral rules across 60 societies with ZERO counter-examples. Alfano et al. (2024) replicated across 256 societies via machine-reading analysis, finding \"evidence for moral universalism.\" This directly supports the convergence thesis.\n- **Fishkin's \"America in One Room\" (2019)**: 22 of 26 extreme partisan positions moved toward center after structured deliberation. Effects lasted 1+ year. Republican deportation support dropped 78.7% → 40.3%. This is not \"averaging out\" — it's preference transformation through exposure to structured reasoning.\n- **Trust → GDP**: Each 15-percentage-point increase in interpersonal trust raises annual GDP growth by ~1 percentage point (Zak & Knack 2001). Nordic countries show 60%+ trust; Latin American countries show under 10%. The transaction sector accounts for 35% of U.S. employment.\n\n### 13.2 Deutsch's Economic Framework (from deep Deutsch research)\n\nDeutsch provides the philosophical foundation missing from Wright's historical narrative:\n\n- **Wealth redefined**: \"The repertoire of physical transformations you are capable of bringing about.\" Wealth is knowledge-dependent, not resource-dependent. Copper was useless rock until metallurgical knowledge existed. Silicon was sand until semiconductor knowledge. Resources are not fixed — they are created by knowledge.\n- **Constructor theory as institutional design**: A constructor is an entity that performs a task while retaining the ability to perform it again. A society is a constructor for cooperation. An institution that destroys its own error-correction capacity (suppresses criticism, punishes dissent) is a *self-consuming constructor* — it performs the task but loses the ability to do it again. Deliberus is designed as a non-self-consuming constructor for collective reasoning.\n- **Payoff matrices are knowledge-dependent**: Deutsch doesn't engage formal game theory, but his framework radically reframes it. What looks like a zero-sum game with fixed payoffs becomes positive-sum when creative option-generation changes the game itself. The Fisher/Ury \"expanding the pie\" is a specific instance of this: decompose positions into interests and the apparent conflict dissolves. Deliberus's claim decomposition IS this process, systematized.\n- **The moral boundary**: For Deutsch, the line between dynamic (progressive) and static (stagnant) societies is willingness to participate in error-correction. Static societies refuse criticism; dynamic societies embrace it. This maps onto Deliberus as a moral commitment: the platform IS the commitment to error-correction in collective reasoning.\n- **\"Deutsch's law\" (from Naval conversations)**: Every interesting problem is soluble. Not because we already have the solution, but because no known law of physics prevents us from finding it. The practical consequence: pessimism about cooperation (\"humanity will never agree on X\") assumes a fixed level of knowledge. Build better epistemic institutions and the level of knowledge changes.\n\n### 13.3 Collective Intelligence Empirics (from collective intelligence research)\n\n- **The c-factor is real**: Woolley et al. (*Science* 2010, n=699) showed a measurable collective intelligence factor that correlates with social sensitivity and equal turn-taking, NOT average individual IQ. Groups with higher c-factors outperform groups of high-IQ individuals with lower social sensitivity.\n- **DeliData's stunning finding (2023)**: 64% of groups found better solutions than individuals. But the remarkable detail: **43.8% of successful groups contained NO individual who had solved the problem alone** — the correct answer emerged purely from group interaction. This is empirical proof that group reasoning produces knowledge no individual possesses.\n- **Landemore's epistemic democracy**: Cognitive diversity, not expert ability, is the key epistemic resource. Random selection (sortition) produces more diverse groups than election. Her 2026 book *Politics without Politicians* argues for full lottocracy. This provides theoretical justification for open-contribution argument graphs over expert-curated platforms.\n- **Benkler's three conditions for peer production**: Modularity, fine granularity, low-cost integration. Deliberus maps cleanly: claims are modules (independently contribute-able), votes are fine-grained (the smallest meaningful contribution), and QBAF is the integration mechanism (automatically computing aggregate argument strength from individual contributions).\n- **Dryzek & List (2003)**: Deliberation can escape Arrow's impossibility theorem by transforming preferences into single-peaked profiles. This is formally significant — it means Deliberus operates at the right layer (preference transformation through reasoning, not just preference aggregation through voting).\n\n### 13.4 Game Theory and Mechanism Design (from game theory research)\n\n- **Axelrod's gap**: In the Evolution of Cooperation tournaments, agents could only cooperate or defect — they couldn't communicate *why*. Adding a reasoning channel (which Deliberus provides) fundamentally changes the game: agents can discover that their payoff matrices were based on misunderstandings.\n- **Supermodular games**: When players' strategies are complements (more cooperation by one increases returns to cooperation by others), the game exhibits strategic complementarities. Novel argument: deliberation graphs are supermodular — each contribution increases the marginal value of others' contributions, because each new claim creates new connection possibilities, fills evidence gaps, and enriches the shared reasoning resource.\n- **The expanding pie IS claim decomposition**: Fisher and Ury's integrative negotiation (decomposing positions into interests to find mutual gains) is structurally identical to Deliberus's claim decomposition. The classic orange example (two people want the same orange — one wants juice, one wants rind) IS contested concept detection applied to negotiation. Making the deeper structure visible reveals the cooperative solution.\n- **Coase theorem for deliberation**: The six barriers to Coasean bargaining (information asymmetry, communication costs, bounded rationality, transaction costs, strategic behavior, cognitive biases) each map to a specific Deliberus mechanism that reduces them. The Cosmos Institute's \"Coasean Bargaining at Scale\" thesis makes this connection explicitly.\n- **Single convergent finding**: All eight game-theory frameworks (Wright, Axelrod, mechanism design, supermodular games, Ostrom, Fisher/Ury, Coase, Romer) converge on the same bottleneck: **not human selfishness, but the cost of mutual understanding.** Reducing that cost is Deliberus's core function.\n\n### 13.5 AI Cooperation Infrastructure (from AI cooperation research)\n\n- **Cooperative AI Foundation funding opportunity**: $15M endowment, board includes Audrey Tang (Taiwan digital minister who championed vTaiwan). Directly funds \"AI for Facilitating Human Cooperation.\" Grants include Procaccia at Harvard ($233K for policy aggregation) and Doshi-Velez ($214K for humanitarian negotiation).\n- **Community Notes at scale**: Reduces reposts by 46%, structural virality by 48.5% (PNAS, n=40,078). But only 29% of fact-checkable tweets get helpful notes, taking 7-70 hours. Validates bridging-based scoring at scale. Misses argument structure — shows that bridge-finding works but is insufficient without decomposition.\n- **Decidim (925K participants, 490 instances, 32 countries)**: Succeeds at participation but has zero argument structure. Polis (10M+ participants, 80% government action rate in Taiwan): succeeds with bridging but has no reasoning decomposition. **No existing system combines persistent graphs + decomposition + bridging detection + progressive disclosure.** This is Deliberus's distinctive gap.\n- **DCI paper validation**: Prakash (2026) showed structured deliberation with 14 typed epistemic acts outperforms unstructured debate on hidden-profile tasks. The typed acts closely mirror Walton's scheme classification — independent convergence on the same architectural insight.\n- **Alignment through structured deliberation**: DeepMind's 2024 debate study (5M model calls, 9 domains) proves stronger debaters lead to higher judge accuracy even for weak judges. If Deliberus can enhance argument quality through scheme classification and CQ generation, it simultaneously serves as \"judge enhancement infrastructure\" for AI alignment.\n\n---\n\n## Sources\n\n### Books and Major Works\n- Wright, R. (1999). *Nonzero: The Logic of Human Destiny*. Pantheon.\n- Deutsch, D. (2011). *The Beginning of Infinity: Explanations That Transform the World*. Viking.\n- Ostrom, E. (1990). *Governing the Commons: The Evolution of Institutions for Collective Action*. Cambridge University Press.\n- Fishkin, J. (2018). *Democracy When the People Are Thinking*. Oxford University Press.\n- Popper, K. (1945/2013). *The Open Society and Its Enemies*. Princeton University Press.\n- Coase, R. (1960). \"The Problem of Social Cost.\" *Journal of Law and Economics*, 3, 1-44.\n- Weber, S. (2004). *The Success of Open Source*. Harvard University Press.\n\n### Academic Papers\n- [Procaccia, Konya et al. (2025). \"Using Collective Dialogues and AI to Find Common Ground Between Israeli and Palestinian Peacebuilders.\" FAccT 2025](https://arxiv.org/abs/2503.01769)\n- [Dafoe, A. et al. (2020). \"Open Problems in Cooperative AI\"](https://arxiv.org/abs/2012.08630)\n- [Klingefjord, O., Lowe, R. & Edelman, J. (2024). \"What are human values, and how do we align AI to them?\"](https://arxiv.org/abs/2404.10636)\n- [Tessler, M. et al. (2024). \"AI can help humans find common ground in democratic deliberation.\" *Science*](https://www.science.org/doi/10.1126/science.adq2852)\n- [Lees, J. & Cikara, M. (2021). \"False Polarization: Cognitive Mechanisms and Potential Solutions.\" *Current Opinion in Psychology*](https://www.sciencedirect.com/science/article/abs/pii/S2352250X21000749)\n- [Blatz, C. & Mercier, B. (2018). \"False Polarization and False Moderation.\" *Social Psychological and Personality Science*](https://journals.sagepub.com/doi/abs/10.1177/1948550617712034)\n- [Prakash, A. (2026). \"From Debate to Deliberation.\" arXiv:2603.11781](https://arxiv.org/abs/2603.11781)\n- [Collective cooperative intelligence. *PNAS* (2024)](https://www.pnas.org/doi/10.1073/pnas.2319948121)\n- Stenseke, J. (2024). \"Morality is Hard: NP-Hardness of Ethical Decision-Making.\"\n- [Twelve Key Findings in Deliberative Democracy Research — American Academy of Arts and Sciences](https://www.amacad.org/publication/daedalus/twelve-key-findings-deliberative-democracy-research)\n\n### Reports and Web Sources\n- [Hidden Tribes: A Study of America's Polarized Landscape — More in Common (2018)](https://hiddentribes.us/)\n- [The Perception Gap — More in Common (2019)](https://perceptiongap.us/)\n- [Cooperative AI Foundation — 2024 Strategy](https://www.cooperativeai.com/post/the-cooperative-ai-foundations-2024-strategy)\n- [Robert Wright TED Talk: \"Progress is not a zero-sum game\"](https://www.ted.com/talks/robert_wright_progress_is_not_a_zero_sum_game)\n- [David Deutsch and Naval Ravikant — Tim Ferriss Show #662](https://tim.blog/2023/03/24/david-deutsch-naval-ravikant-transcript/)\n- [Anti-rival good — Wikipedia](https://en.wikipedia.org/wiki/Anti-rival_good)\n- [Ostrom — \"Governing the Commons\" Summary — Beyond Intractability](https://www.beyondintractability.org/bksum/ostrom-governing)\n- [David Deutsch: Knowledge Creation and The Human Race — Naval](https://nav.al/david-deutsch)\n- [Nonzero — Wikipedia](https://en.wikipedia.org/wiki/Nonzero:_The_Logic_of_Human_Destiny)\n\n### Companion Research Documents (produced by parallel research agents, April 6, 2026)\n- [docs/research/david-deutsch-deep-research-non-zero-sum-economics.md](david-deutsch-deep-research-non-zero-sum-economics.md) — 453 lines, constructor theory, wealth redefinition, error-correcting institutions, Naval conversations\n- [docs/research/non-zero-sum-game-theory-and-cooperation-economics.md](non-zero-sum-game-theory-and-cooperation-economics.md) — 485 lines, Wright, Axelrod, mechanism design, supermodular games, Ostrom, Coase\n- [docs/research/collective-intelligence-and-epistemic-democracy.md](collective-intelligence-and-epistemic-democracy.md) — ~400 lines, Malone, Landemore, Page, citizens' assemblies, DeliData\n- [docs/research/trust-economics-and-false-polarization.md](trust-economics-and-false-polarization.md) — 361 lines, Fukuyama, false polarization quantified, identity-protective cognition, moral universalism\n- [docs/research/ai-augmented-cooperation-infrastructure.md](ai-augmented-cooperation-infrastructure.md) — ~400 lines, Cooperative AI Foundation, Procaccia/Konya, Community Notes, digital public goods\n\n### Internal Deliberus References\n- [docs/civilizational-vision.md](../civilizational-vision.md) — trust transformation, resource allocation, AI alignment\n- [docs/convergence.md](../convergence.md) — the convergence thesis, MGE empirical support, Lean analogy\n- [docs/bridging.md](../bridging.md) — bridging arguments, the \"I can't dismiss this\" moment, the formula\n- [docs/depth.md](../depth.md) — No Copout Axioms, universal acid, self-similar decomposition\n- [docs/vision.md](../vision.md) — core intent, key tensions, the destination\n- [docs/analysis-and-attunement.md](../analysis-and-attunement.md) — the core dialectic\n- [docs/worldview-lenses.md](../worldview-lenses.md) — the worldview filter\n- [docs/research/david-deutsch-beginning-of-infinity-and-deliberus.md](david-deutsch-beginning-of-infinity-and-deliberus.md) — Deutsch's epistemological contributions (session 10)\n- [docs/research/zizek-schmachtenberger-connections.md](zizek-schmachtenberger-connections.md) — anti-rivalrous dynamics, metacrisis, cynical ideology\n- [docs/research/lean-deliberus-analogies.md](lean-deliberus-analogies.md) — sorry-driven blueprints, `@[simp]` flywheel, trustless collaboration\n- [docs/research/habermas-machine.md](habermas-machine.md) — AI-mediated consensus, Schulze voting\n- [docs/research/prediction-markets-argumentation.md](prediction-markets-argumentation.md) — LMSR, futarchy, argument markets\n- [docs/research/epistemic-gamification.md](epistemic-gamification.md) — calibration, delta system, anti-gaming\n- [docs/research/simplenote-archive-analysis.md](simplenote-archive-analysis.md) — earliest formulations (2011-2012)\n- [docs/research/chat-logs-analysis.md](chat-logs-analysis.md) — hybrid AGI, ethical filters, resource-based economy connection\n- [docs/research/governance-and-financing.md](governance-and-financing.md) — organizational models, funding paths\n"}